Improving quality of service for resource allocation in double auction mechanism using modified heuristic techniques

Abstract

Cloud computing makes use of a virtual technology to map the physical resources to the virtual machines to carry out the tasks of the users through resource based scheduling. It has been showed that in many researches, the distribution of resources in cloud computing is a problem that is NP based. There are multiple targets in the allocation of resources based on the tasks to be performed. In solving this, both the domestic and foreign scholars have been introducing many artificial intelligence algorithms in the resource distribution and are achieving significant results. In this, a resource allocation mechanism has the aim of optimizing the resource utilization for satisfying the demands of the customers. An Auction is a market mechanism which is used widely for allocating and selling resources. In case of a double sided auction the customers and the providers submit the bids and the asks respectively. A market based pricing algorithm is used. The trust model is based on the double auction mechanism where multiple parties trade cloud resource based on each other needs. Trading takes place immediately when there is a match or compatibility between the bids and the asks. Resource provisioning also includes identifying the cloud providers, virtual machine scheduling and utilizing the resources in an efficient manner on the basis of the pricing scheme and the management of resources and workload. Various metaheuristic algorithms are available with novel variations and are suggested for the allocation of resources in many fields. In this research, cloud auction method with a close to optimal resource allocation using Branch and Bound (BandB) method, Backtracking algorithm and Greedy Search algorithm is performed. The BandB framework is a widelyand fundamentally used method in the production of the exact solutions to that of the problems of NP-hard optimization. The Greedy Search is based on the sequential application of auction algorithm. In each step a resource is allocated to a target and greediness improves the time